W. A. Davis, A. J. Daisii and G. C. Sawyer 273 



at 90 volts drop, the last three contacts giving 10 amps, at 80, 70 and 

 60 volts drop. This was made for us by Messrs Tyler and Freeman, 

 Loudon. 



Experience has shown that by taking eight to ten readings it is 

 possible with our instrument to obtain values which have a probable 

 error considerably less than ± 0-005°. Duplicate sets of readings with 

 a clear solution generally agree to within 0-002°. 



Probable error of the Analyses and Methods of Sanifling. 



A. Methods of analysis. To ascertain whether any inversion of 

 cane sugar is caused during the evaporation in vacuo and the subsequent 

 treatment of the alcoholic extract, and also the degree of accuracy with 

 which cane sugar can be estimated in such solutions, the original extract 

 from a picking of mangold leaves having a volume of about 3 litres was 

 made up exactly to 4 litres. It was then divided into two halves, each 

 exactly of 2 litres, and to one of the halves 5-00 grms. of pure saccharose 

 was added, to the other nothing. Each half was then evaporated in 

 vacuo as usual and subjected to the ordinary processes of analysis. 

 As a result it was found that the difference between the average values 

 in the two cases, by the reduction -process, using both the iuvertase and 

 citric acid methods of inversion, corresponded with 4-98 grms. instead 

 of the 5-00 grms. actually added. The increase of polarisation in the 

 direct solution due to the added sugar gave a slightly higher value, 

 5-22 grms. saccharose, whilst the results calculated from the change of 

 polarisation on inversion gave for the added sugar : 



1. Invertase 5-80 ) 



r,- ■ -1 /.,,. r average = 5-98 grms. 



2. Citric acid 6-16 / ^ ^ 



The results obtained by the change of polarisation on inversion are 

 therefore nearly 20 per cent. high. Throughout our work we have 

 observed a similar divergence between the results obtained by the 

 reduction and polarisation methods; the explanation of this is given 

 later (see p. 329). Consequently in discussing the variation of the sugars 

 throughout the day we have used the reduction values only and ignored 

 the polarisation results, which are undoubtedly high in the majority of 

 cases and especially so in the case of the stalks and mid-ribs. The value 

 4-98 obtained for saccharose by the reduction method as compared with 

 the 5-00 grms. actually added shows that no inversion or loss of cane 

 sugar is to be feared under our conditions of working and that the reduc- 

 tion method is quite trustworthy. 



19—2 



